Duplexed CeTEAM drug biosensors reveal determinants of PARP inhibitor selectivity

Abstract

PARP inhibitors (PARPi) predominantly targeting PARP1 and PARP2 have revolutionized cancer therapy by selectively killing cancer cells with defective DNA repair. However, achieving PARP1 or PARP2-selective inhibitors is difficult due to their close structural homology. Selectivity profiling is typically done with purified proteins, but these lack the complexity of intracellular environments and could therefore be inaccurate. Here, we duplex PARP1 L713F-GFP and PARP2 L269A-mCherry CeTEAM drug biosensors to systematically characterize binding and cell cycle alterations of 27 PARPi at the single cell level. Our results reveal that most PARPi are generally equipotent for both PARPs, including the next-generation drug, senaparib. However, benzimidazole carboxamide (niraparib) derivatives demonstrated PARP1-selective tendencies, while pthalazinones (olaparib) favored PARP2. AZD5305, a reported PARP1-selective inhibitor with characteristics of both series, was the exception and appears [~]1600-fold more potent towards PARP1. In agreement with current understanding, we see that PARP trapping phenotypes positively correlate with PARP1/2 binding potency, while some potent binders, such as veliparib, did not - likely reflecting their allosteric influence on DNA retention. We also assessed the effect of the PARP1/2 active site component, HPF1, on intracellular PARPi binding and see that HPF1 depletion elicits slight deviations in apparent binding potency, while contributing additively to PARP-DNA trapping phenotypes. The PARP1/2 CeTEAM platform thus provides a structural roadmap for the development of selective PARPi and should facilitate the discovery of better targeted therapies. Furthermore, our results highlight that multiplexing CeTEAM biosensors and layered genetic perturbations can systematically profile determinants of intracellular drug selectivity.